LITMIR.BIZ

      Preface

      P.1. The book series

      This book constitutes the first volume of the set of books entitled “IsoGeometric Analysis (IGA) tools for optimization applications in structural mechanics”. The objective of the series is to present, in a comprehensive and detailed manner, some advanced modeling and numerical strategies, recently developed in the context of IGA, that provide strong benefits not only for direct simulations but also for the resolution of optimization problems in structural mechanics. The IGA paradigm was originally introduced by Hughes and co-workers in 2005 (Hughes et al. 2005) to reunify the fields of computer-aided design (CAD) and finite element analysis (FEA). The core idea is to resort to the same higher-order and smooth spline bases for the representation of the geometry in CAD as well as for the approximation of solutions fields in FEA. The use of such families of functions quickly made IGA highly attractive for two main reasons: first, a common geometrical model can be used by both the designers and analysts; second, an increased per-degree-of-freedom (per-DOF) accuracy can be reached in comparison to standard finite element methods (FEM). This technology is thus often seen as a high-performance computational tool.

Beyond its undeniable superior analysis properties, IGA is also highly relevant for addressing the higher-level problems of optimization (also characterized as inverse problems) that enable us to question and adapt a numerical model with regard to key features. Indeed, IGA provides a natural regularization framework for such problems since it allows us to look for the solution in more regular approximation subspaces. Following this mindset, many progresses can be reported in the current literature, with the common goal of further consolidating IGA by demonstrating its performance for optimization-type applications. This book series is meant to be part of this attempt.

P.2. The present volume

      This first volume tackles design optimization that is essential to reduce the usual long trial-and-error learning process during the product development in industry. More precisely, we focus on structural shape optimization, which aims at finding the optimal shape of a given topology of a structure with respect to a certain objective, such as minimal mass and maximal rigidity. Shape optimization undoubtedly constitutes one of the most valuable applications for IGA as the latter combines high-quality geometric representations and efficient analysis capabilities into one single model. In this book, we seek to develop a full analysis-to-optimization framework that is applicable to complex structures in order to tend to real-world (industrial) applications.

      From an analysis point of view, this forces us to face two very important challenges encountered in the field today, namely (i) the simple local enrichment of rigid tensor-product spline models (which may also go with trimming when introducing an arbitrary local region within a spline patch) and (ii) the efficient analysis of large multipatch structures. The methods proposed to answer these issues are quite original for the domain and are based on developments in which the French community is particularly active. More precisely, the strategies rely on domain coupling and, in particular: (i) global/local non-invasive coupling algorithms that enable us to locally enrich a global model without altering its corresponding numerical operators and (ii) parallel domain decomposition solvers that allow us to separate the computational resources between several subdomains.

In a next step, the developed direct solvers are finely integrated into the shape optimization loop for full robustness and efficiency, and a geometrical modeling, based on embedded entities, is formulated to arrive at what we may refer to as optimization-suitable models, i.e. models that are ready for natural isogeometric shape optimization regardless of their complexity. The resulting framework is very robust and generic; it merges all the attractive features of IGA (design–analysis link, numerical efficiency, natural regularization, etc.) and thus brings the possibility of exploring new types of design. The performance of the methodology is demonstrated in the context of elastic solid and thin shell structures (with the Kirchhoff–Love theory). In particular, we consider as a guiding application the optimal design of stiffened structures, which are ubiquitous in aeronautics. Today, designs of wings and fuselages contain a lot of straight and aligned stiffeners uniformly distributed. However, it seems that defining curvilinear stiffeners instead of straight ones can further improve the mechanical behavior of the overall stiffened structures. Figure P.1 highlights what could be, in the future, the aerostructures based on free-form stiffeners. The developed framework contributes to reach such innovative designs.